Prognostic value of tumor stromal collagen features in patients with hepatocellular carcinoma revealed by second-harmonic generation microscopy

Hepatocellular carcinoma (HCC) is the sixth most common cancer worldwide. The search for new biomarkers that predict the outcome of HCC patients is ongoing. We propose the second harmonic generation–based quantitative assessment approach to evaluate the prognostic value of tumor stromal collagen in HCC. We evaluated tumor stromal collagen in paraffin-embedded specimens from 109 HCC patients by second-harmonic generation imaging. The parameters and quantitative assessment of collagen were obtained using a fiber network extraction algorithm. The relationships between collagen features and clinical pathological features and overall survival were statistically analyzed. Among the collagen features, some parameters of aggregated collagen correlated well with clinical pathological features, especially the aggregated collagen cross-linked density. Cross-linked collagen fibers form a fiber network in moderately and poor differentiated HCCs. Kaplan–Meier analyses and the multivariate Cox proportional hazard model showed that high aggregated collagen cross-linked density was associated with poor overall survival. The chi-squared test showed that aggregated cross-link density was significantly associated with histological grade and tumor recurrence. Our results indicate the prognostic value of the quantitative evaluation of tumor stromal collagen using second harmonic generation imaging of patients with HCC. •Collagen plays an important role in the tumor microenvironment.•SHG was used to quantitatively evaluate the prognostic value of tumor stromal collagen of HCC.•Collagen features were determined using a fiber-network extraction algorithm.•High aggregated collagen cross-link density in HCC was associated with poor overall survival.